Uncertainty monitoring for human-like behavior regulation of trajectory planning

Abstract

A method for monitoring uncertainty in human-like behavior regulation of trajectory planning, comprising: retrieving a map and agent information of a current driving state of a host motor vehicle operating autonomously; separating uncertainty conditions affecting a trajectory of the host motor vehicle into expected uncertainty and unexpected uncertainty; calculating the expected uncertainty in a first operational branch by forming an attention zone from an identified portion of a lane that can potentially conflict with a planned route of the host motor vehicle; determining the unexpected uncertainty in a second operational branch by calculating an anomaly score of any other vehicle located in a surrounding area of the host motor vehicle in the lane that can potentially conflict with the planned route of the host motor vehicle; and, if the unexpected uncertainty reaches or exceeds a predetermined threshold, adjusting a trajectory operational signal determined for the expected uncertainty.

Description

Technical Field

[0001] This disclosure relates to systems and methods for generating and regulating the trajectories of autonomous vehicles. Background Technology

[0002] Autonomous vehicles handle uncertainty in their environment to make decisions and adjust their actions. Existing autonomous vehicle handling methods do not operate on human-like systems and approaches, thus lacking compatibility with human drivers and passengers.

[0003] Many traditional autonomous driving methods monitor and process uncertainty; however, these methods do not characterize uncertainty based on the brain's neural regulatory principles. Therefore, traditional autonomous driving methods lack human-compatible adaptive processes for behavioral and computational adjustments.

[0004] Therefore, although current autonomous vehicle trajectory calculation systems and methods have achieved their intended purpose, a new and improved approach is still needed to monitor the uncertainty of human-like behavior regulation in trajectory planning. Summary of the Invention

[0005] According to several aspects, a method for monitoring uncertainties in humanoid behavior regulation for trajectory planning includes: retrieving a map and agent information of the current driving state of an autonomously operated master motor vehicle; classifying uncertainties affecting the trajectory of the master motor vehicle into expected uncertainties and unexpected uncertainties; calculating expected uncertainties in a first operating branch by forming an attention zone based on a lane identification portion, the lane identification portion of which may potentially conflict with the planned route of the master motor vehicle; determining unexpected uncertainties in a second operating branch by calculating the anomaly score of any other vehicle in the area surrounding the master motor vehicle located in a lane that may potentially conflict with the planned route of the master motor vehicle; and adjusting the trajectory operation signal determined for the expected uncertainties if the unexpected uncertainties reach or exceed a predetermined threshold.

[0006] In another aspect of this disclosure, the method further includes retrieving map and agent information of the current driving state, including representing lane information as coordinates, heading, and speed limit.

[0007] In another aspect of this disclosure, the method further includes: expressing the coordinates, heading, and speed of the agent, including other vehicles in the area surrounding the main motor vehicle, during the retrieval of map and agent information for the current driving state.

[0008] In another aspect of this disclosure, the method further includes using an attention zone to filter out agents of any vehicle in the surrounding area of ​​the primary motor vehicle, including the attention zone, such that the filtered-out agents are not used for further processing.

[0009] In another aspect of this disclosure, the method further includes: generating potential trajectory branches from the trajectory of the main motor vehicle.

[0010] In another aspect of this disclosure, the method further includes selecting the optimal or "best" trajectory branch from the potential trajectory branches to be executed by the main motor vehicle.

[0011] In another aspect of this disclosure, the method further includes comparing the speeds and headings of other vehicles in the area surrounding the main motor vehicle with the individually expected speeds and headings of other vehicles at their current lane positions.

[0012] In another aspect of this disclosure, the method further includes obtaining a summary score of unexpected uncertainty by averaging the anomalous scores of any other vehicles.

[0013] In another aspect of this disclosure, the method further includes: generating a first adjustment signal, the first adjustment signal being applied to temporarily disable attention zone filtering for any other vehicle during the adjustment trajectory operation signal.

[0014] In another aspect of this disclosure, the method further includes: executing a second operation branch in parallel with the first operation branch.

[0015] According to several aspects, a method for monitoring uncertainties in humanoid behavior regulation for trajectory planning includes: retrieving a map and agent information of the current driving state of an autonomously operated master motor vehicle; classifying uncertainties affecting the trajectory of the master motor vehicle into anticipated uncertainties and unexpected uncertainties; determining anticipated uncertainties by forming attention zones based on lane identification portions that may potentially conflict with the planned route of the master motor vehicle; setting a predetermined threshold such that when the level of unexpected uncertainties is below the predetermined threshold, the attention zones are used only for computational savings; and applying the attention zones to reduce the computational cost required to make trajectory decisions, wherein assumptions about vehicles within each attention zone that defines a high attention zone are used to determine when to perform a maneuver, and wherein assumptions about vehicles outside the attention zones are not calculated.

[0016] In another aspect of this disclosure, the method further includes: representing lane information as coordinates, heading, and speed limits during the retrieval of map and agent information for the current driving state; and expressing the coordinates, heading, and speed of agents, including other vehicles in the area surrounding the main motor vehicle.

[0017] In another aspect of this disclosure, the method further includes: using maps and agent information to represent the predicted headings of other vehicles in the area surrounding the primary motor vehicle on the available road path by means of the positions and angles indicated by arrows.

[0018] In another aspect of this disclosure, the method further includes: checking a plurality of planned points in front of the main vehicle within a radius around the main vehicle.

[0019] In another aspect of this disclosure, the method further includes: assessing unexpected uncertainties relative to time; and changing the decay constant to adjust a predetermined threshold over time.

[0020] In another aspect of this disclosure, the method further includes: projecting multiple points defining a planned route for the main motor vehicle; and calculating an angle θ between one of the arrows and one of the multiple points.

[0021] In another aspect of this disclosure, the method further includes: drawing a box around one of the arrows, and including one of the arrows and data associated with one of the arrows in one of the attention areas if the angle θ is less than or equal to about 20 degrees, indicating the path of one of the arrows and the potential intersection of one of the points.

[0022] According to several aspects, a system for monitoring uncertainties in human-like behavior regulation during trajectory planning includes a map and agent information defining the current driving state of an autonomous driving vehicle. Uncertainty conditions affecting the trajectory of the driving vehicle can be categorized into anticipated uncertainty and unexpected uncertainty. The anticipated uncertainty defining the attention zone is formed based on lane identification portions that may potentially conflict with the planned route of the driving vehicle. The unexpected uncertainty defines anomaly scores calculated for any other vehicles in the area surrounding the driving vehicle located in lanes that may potentially conflict with the planned route of the driving vehicle. A trajectory operation signal is determined for the anticipated uncertainty. In another aspect of this disclosure, an attention zone defining filter operates to filter out agents, including any vehicles in the area surrounding the driving vehicle outside the attention zone, thereby avoiding further processing and elimination of the filtered-out agents.

[0023] In another aspect of this disclosure, an adjustment signal is generated, which adjusts the trajectory operation signal if an unexpected uncertainty reaches or exceeds a predetermined threshold.

[0024] Further areas of application will become apparent from the description provided herein. It should be understood that the descriptions and specific examples are intended for illustrative purposes only and are not intended to limit the scope of this disclosure. Attached Figure Description

[0025] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of this disclosure in any way.

[0026] Figure 1It is a graphical flowchart of method steps according to an exemplary aspect, the method steps being used to perform a method for monitoring the uncertainty of humanoid behavior regulation in trajectory planning;

[0027] Figure 2 This is a top view of an exemplary four-way intersection, depicting the area of ​​attention under the anticipated uncertainty conditions of this disclosure;

[0028] Figure 3 It is a top view of an exemplary straight road, depicting the area of ​​attention of this disclosure;

[0029] Figure 4 It is a schematic diagram of the planned route of the main motor vehicle and the potential intersections with other vehicles;

[0030] Figure 5 It is by Figure 2 Modified top view to depict the area of ​​attention for unexpected uncertainty conditions in this disclosure;

[0031] Figure 6 It is by Figure 3 Modified top view to depict the area of ​​attention for unexpected uncertainty conditions in this disclosure;

[0032] Figure 7 It is a graph representing the curve of unexpected uncertainty covered by a predetermined threshold line;

[0033] Figure 8 It is a graph representing the unexpected uncertainty curve covered by a predetermined static threshold line generated using a zero-value decay constant;

[0034] Figure 9 It is a graph representing the curve of unexpected uncertainty covered by the changing threshold line;

[0035] Figure 10 It is a graph representing the unexpected uncertainty curve covered by the threshold line of the tight arbitrary external uncertainty curve; and,

[0036] Figure 11 It is a graph representing the unexpected uncertainty curve due to the fact that the threshold line is not covered because the decay constant is equal to 1. Detailed Implementation

[0037] The following description is exemplary in nature and is not intended to limit this disclosure, application, or use.

[0038] refer to Figure 1The system and method for monitoring uncertainties in the human-like behavior regulation of trajectory planning 10 utilize uncertainty calculations to influence trajectory output to control the operation of an autonomous main motor vehicle. In the initial step 12, the system and method for monitoring uncertainties in the human-like behavior regulation of trajectory planning 10 first retrieve map and agent information for the current driving state, including lane information represented as coordinates, heading, and speed limits, as well as the coordinates, heading, and speed of agents, including other vehicles around the main motor vehicle. Uncertainty conditions affecting the trajectory of the main motor vehicle are categorized into anticipated uncertainties and unexpected uncertainties. Then, a bifurcation operation is performed in parallel to independently resolve anticipated uncertainties and identify whether unexpected uncertainties justify modifying the trajectory of the main motor vehicle.

[0039] The anticipated uncertainty is resolved in the first operation branch 14, and the unexpected uncertainty is resolved in the parallel second operation branch 16. If the unexpected uncertainty determined during the parallel second operation branch 16 reaches or exceeds a predetermined threshold, the unexpected uncertainty can be used to adjust the trajectory operation signal generated in response to the anticipated uncertainty determined in the first operation branch 14.

[0040] During the first operational branch 14, in the first stage 18, anticipated uncertainty is calculated by forming an attention zone based on the identified portion of the lane, which may potentially conflict with the planned route of the main motor vehicle. In the second stage 20, the attention zone is then used to filter out agents, including other vehicles and objects outside the attention zone, so that the filtered agents are not used for further processing. In the third stage 22, potential trajectory branches are generated. In the subsequent fourth stage 24, the optimal or "best" trajectory branch is selected from the potential trajectory branches for execution by the main motor vehicle or as a suggestion to the operator of the main motor vehicle. The method used for branch generation and selection is not limited to the systems and methods of this disclosure; the only requirement is that the trajectory generation has the ability to vary speed and latency.

[0041] In parallel with the phases executed during the first operational branch 14, during the parallel second operational branch 16, unexpected uncertainty is calculated in the fifth stage 26. Unexpected uncertainty can be used to adjust the trajectory generation and selection process that occurs during the first operational branch 14. Unexpected uncertainty is determined by calculating anomaly scores for each other vehicle in the area surrounding the primary motor vehicle, by comparing the speed and heading of one or more other vehicles with the individually expected speed and heading of the current lane position of other vehicles. A summed score of unexpected uncertainty 28 is then obtained by averaging the anomaly scores of all other vehicles. Using unexpected uncertainty 28, in the sixth stage 30, the calculated unexpected uncertainty 28 is compared with a first threshold 32, and if unexpected uncertainty 28 exceeds the first threshold 32, a first adjustment signal 34 is generated to modify the second stage 20, thereby temporarily disabling attention zone filtering for other vehicles.

[0042] If the calculated unexpected uncertainty 28 does not exceed the first threshold 32, then in the seventh stage 36, based on the expected slower trajectories and longer waiting times of other vehicles behind stop signs and other temporary roadblocks, an analysis of the potential trajectory branches initially identified in the third stage 22 is performed. The sum of trajectory branches 38 is inversely proportional to the calculated unexpected uncertainty 28. In the seventh stage 36, the sum of calculated trajectory branches 38 is compared with a second threshold 40, and if the sum of calculated trajectory branches 38 exceeds the second threshold 40, a second adjustment signal 42 is generated to modify the third stage 22 to adjust the available trajectory branches.

[0043] In stage 8, 44, it is determined whether the optimal or “best” trajectory branch from the potential trajectory branches in stage 4, 24, needs adjustment to change the selected trajectory taken by the main motor vehicle. If a third threshold 46 is exceeded, a risk-avoidance strategy (such as stopping the main motor vehicle) is selected in the final trajectory selection, and when the unexpected uncertainty exceeds the third threshold 46, a third adjustment signal 48 is sent to modify stage 4, 24.

[0044] The impact of different types of uncertainty on behavior and computation has been proposed, with each component corresponding to different brain functions and circuits. Anticipatory uncertainty is modeled after the neuromodulator acetylcholine, and may be constrained by attentional regions. Unexpected uncertainty is modeled after the neuromodulator norepinephrine, and can be constrained using anomaly scores, such as heading-based or velocity-based anomalies. Trajectories can be described along two axes of opposite behavior. One is reward-seeking behavior, constrained by preferred velocity and response to nearby agents, and modeled by the neuromodulator dopamine. The other is risk aversion, constrained by a willingness to wait for uncertainty to decrease before proceeding with an intentional move, as modeled by the neuromodulator serotonin.

[0045] The primary vehicle can use attention zones to treat other agents differently. For example, the hypothetical trajectory of a vehicle within a high attention zone can receive more weight, while agents completely outside the zone can be ignored entirely.

[0046] refer to Figure 2 and Figure 3 And refer to again Figure 1 Autonomous or primary motor vehicle 50 can use attention zones to reduce the computational burden required to make trajectory decisions. Assumptions about vehicles within high attention zones can be used to determine when it is safe to perform maneuvers, without needing to compute assumptions about vehicles outside the attention zone. This approach reduces the processing and information required for trajectory decisions. By filtering vehicles outside the attention zone, the computational savings allow for more processing of safety-based features, directing attention towards processes that ultimately maximize the safety of the primary motor vehicle 50. Using knowledge of anticipated traffic flow, attention zones can be formed by predicting potential collision areas if the vehicle follows the traffic flow.

[0047] For example, such as Figure 2 As illustrated more specifically, if a primary motor vehicle 50 is stopped at a four-way intersection with traffic lights and intends to turn right, the important area of ​​attention is the future traffic flow from left to right, as well as the flow of vehicles from the right that may be making a U-turn.

[0048] Continue to refer to Figure 2 This includes a high attention zone 52 associated with left-to-right traffic flow and a medium attention zone 54 associated with traffic flow from right-hand vehicles that may be making a U-turn within the first map 56, corresponding to potential collisions when driving the main motor vehicle 50 along the intended trajectory 58 that defines a right turn. Vehicles located outside the high attention zone 52 and medium attention zone 54 can receive less computation to reduce computational costs.

[0049] For more specific reference Figure 3A primary motor vehicle 50 traveling on a straight highway has a high collision potential directly ahead and a moderate collision potential with vehicles in adjacent lanes. Therefore, the attention zones are represented as: a high attention zone 60 directly in front of the primary motor vehicle 50, a first medium attention zone 62 on the left side of the primary motor vehicle 50, and a second medium attention zone 64 on the right side. These attention zones in the second map 66 correspond to the potential collisions when the primary motor vehicle 50 travels along the expected straight trajectory 68. Vehicles located outside the high attention zone 60 and outside the first and second medium attention zones 62 and 64 can receive less computation to reduce computational costs.

[0050] The presence of unexpected uncertainty affects anticipated uncertainty by prohibiting the use of attention zones, as agents should pay attention to all zones under abnormal circumstances.

[0051] refer to Figure 4 And refer to again Figures 1 to 3 The calculation of anticipated uncertainty generates a set of attention zones. The main motor vehicle 50 has a current position 70, and a set of its future planned points 72 are provided as input. Using map data, all predicted headings of other vehicles or available road paths in the surrounding area are represented by the positions and angles indicated by arrows. The number of planned points 72 ahead of the main motor vehicle 50 (defined as "num_points") and the radius "radius" around the main motor vehicle 50 to find the arrows are configurable parameters. The following pseudocode calculates the attention zones.

[0052] For each time step: For each point in `num_points` ahead of the current position 70 of the current main vehicle 50, for each arrow (expected heading in the surrounding area) within the radius of the main vehicle 50, if the angle θ between the arrow and the point is sufficiently small, for example, angle θ is less than or equal to approximately 20 degrees, indicating a potential intersection of the arrow path and at least one planned point 72 ahead of the main vehicle 50, then a box is drawn around the arrow, and the arrow and its data are included in the "attention area". If the angle θ between the arrow and the point is large, for example, angle θ is greater than approximately 20 degrees, then no box is drawn around the arrow, and the arrow and its data are not included in the "attention area". Figure 4 In the example provided, a box will be drawn around arrow 74, but not around arrow 76.

[0053] General Reference Figure 5 and Figure 6 And refer again Figures 1 to 4Unexpected uncertainties arise when a vehicle's behavior does not match its intended behavior. For example, a vehicle that violates traffic rules may deviate from its intended speed or direction. When unexpected uncertainties increase due to this anomaly, the main vehicle 50 may decide to change its current maneuvering as appropriate.

[0054] For details, please refer to the following: Figure 5 Increased uncertainty, such as abnormal vehicle 78 not following the expected course and traveling on a course that may intersect with the current course of main vehicle 50, can trigger different behaviors in the decision-making of main vehicle 50. For example, using course-based anomalies, if an agent of the abnormal location (such as abnormal vehicle 78) exists at intersection 80 affecting the course of main vehicle 50, main vehicle 50 may decide to stop.

[0055] For details, please refer to the following: Figure 6 The stalled vehicle 82 exists in the direct path of the main motor vehicle 50. Using speed-based anomalies, the main motor vehicle 50, upon encountering the stalled vehicle 82 with a significantly different speed, can decide to change its driving lane by employing lane change maneuvering path 84 from the current lane 86 to the expected lane 88 (if available).

[0056] refer to Figure 7 To determine when to use the attention zone, a predetermined threshold can be set, so that the attention zone is only used for computational savings when the level of unexpected uncertainty is below the predetermined threshold. The result of using such a predetermined threshold is that computational savings can be achieved in critical situations without sacrificing accuracy. (Reference) Figure 7 Chart 90 represents a series of unexpected uncertainty values ​​92 compared to time period 94. An unexpected uncertainty curve 96 is generated using, for example, unexpected uncertainty values ​​derived from a first instance where vehicle 98 initially overtakes and then overtakes ahead of main vehicle 50. Unexpected uncertainty data 100 from this maneuver produces a first peak 102 in unexpected uncertainty curve 96. In a second instance, vehicle 104 unexpectedly changes lanes, potentially affecting the travel path of main vehicle 50. Unexpected uncertainty data 106 from this maneuver produces a second peak 108 in unexpected uncertainty curve 96. An exemplary predetermined threshold 110 applied to the data in Chart 90 indicates that both the first peak 102 and the second peak 108 exceed the predetermined threshold 110; therefore, since the attention zone will only be used to calculate savings when the unexpected uncertainty level is below the predetermined threshold 110, the attention zone will not be used to generate the conditions for the first peak 102 and the second peak 108.

[0057] General Reference Figures 8 to 11 And refer to again Figure 7Figures 112 through 118 compare the relationship between unexpected uncertainty and time, and the effect of time-varying τ on threshold adjustment. The choice of threshold used for calculation may require manual adjustment based on specific circumstances. To avoid this, the threshold can be automatically adjusted using the following equation based on the background level of uncertainty:

[0058] Equation 1:

[0059] U 意外 =α(h 自 -h 车道 )+(1–α)(v 自 -v 车道 )

[0060] Equation 2:

[0061] τ(dθ / dt)=-(θ–U 意外 )

[0062] The equation symbols are restricted as follows:

[0063] U: Uncertainty

[0064] h: heading

[0065] v: speed

[0066] τ: Attenuation constant

[0067] α: Weight of heading difference relative to speed difference

[0068] θ: Threshold for opening or closing the attention zone

[0069] The attenuation constant τ determines the speed at which the threshold adapts to the background level of uncertainty.

[0070] For more specific reference Figure 8 Figure 112 compares the relationship between unexpected uncertainty 120 and time 122. For the variation values ​​present in the unexpected uncertainty curve 124, the baseline threshold 126 is static and remains at its initially specified value when τ = 0.

[0071] For more specific reference Figure 9 Figure 114 compares the relationship between unexpected uncertainty 128 and time 130. For the same change value of unexpected uncertainty curve 124, as τ increases above 0, it is related to... Figure 8 Compared to the baseline threshold of 126, threshold 132 shows a change.

[0072] For more specific reference Figure 10 Figure 116 compares the relationship between unexpected uncertainty 134 and time 136. For the same change value of unexpected uncertainty curve 124, as τ further increases and approaches 1, it is related to... Figure 9 Compared to threshold 132, threshold 138 shows a closer tracking of the changes in the value of the unexpected uncertainty curve 124.

[0073] For more specific reference Figure 11 Figure 118 compares the relationship between unexpected uncertainty 140 and time 142. For the same change value of unexpected uncertainty curve 124, when τ = 1, the baseline instantaneously follows the uncertainty level, making it unnecessary to use a threshold. Therefore, based on Figures 8 to 11 The optimal value of τ exists between 0 and 1.

[0074] The trajectory generation method of the system and method disclosed herein for monitoring the uncertainty of human-like behavior regulation in trajectory planning 10 is not limited to a specific implementation, but consists of a set of seemingly plausible trajectories with a defined default speed. When the main motor vehicle 50 encounters a temporary obstacle (such as a stop sign, a construction zone, or a stalled vehicle), the generated trajectory causes the vehicle to stop for a defined waiting time. The regulation of trajectory generation alters the parameters of the default speed and the waiting time. The default speed increases inversely with the unexpected uncertainty, and the waiting time increases directly with the unexpected uncertainty, as described in equations 3 and 4 below:

[0075] Equation 3:

[0076] v 默认 =β / U 意外

[0077] Equation 4:

[0078] t 等待 =γU 意外

[0079] Where v 默认 This is the default speed, t 等待 This is the waiting time. β and γ are constant values ​​>1.0, manually adjusted according to driving conditions and regular traffic laws.

[0080] Just as with the adjustment of waiting time, the selection of the final trajectory is also adjusted towards a risk-averse selection. For example, one method of trajectory generation may include a normal trajectory and a fault-protected trajectory consisting of a sudden stop. If the unexpected uncertainty exceeds a predetermined threshold, the fault-protected trajectory will be selected.

[0081] The system and method disclosed herein for monitoring uncertainties in the human-like behavior adjustment of trajectory planning 10 offer several advantages. These advantages include autonomous vehicles that adaptively reduce computation and intelligently adjust driving behavior by simulating human cognition. Vehicle safety is maintained by adjusting computational load according to the level of uncertainty, reducing computation only in low-risk, low-uncertainty situations. By acting in a human-like manner, human drivers in the vicinity of the autonomous vehicle are better able to predict and anticipate potential movements of the autonomous vehicle, enabling smoother driving in traffic situations involving both humans and autonomous drivers, thereby reducing the risk of collisions. The system of this disclosure is also compatible with semi-autonomous applications, such as driver assistance, where the system can be used to provide attentional cues to occupants / operators or generate trajectories for emergency maneuvers.

[0082] A system and method are provided to monitor uncertainties in trajectory planning for autonomous vehicles, thereby reducing path planning computation and adjusting driving behavior. The uncertainty measurement applied in the system and method of this disclosure is based on neural regulatory mechanisms of human cognition. This method leads to a more human-compatible adaptive process in behavioral and computational adjustments.

[0083] The description in this disclosure is exemplary in nature only, and any changes that do not depart from the spirit and scope of this disclosure are also intended to be within the scope of this disclosure. Such changes should not be considered as departing from the spirit and scope of this disclosure.

Claims

1. A method for monitoring uncertainty in human-like behavior regulation during trajectory planning, comprising: Retrieve map information and agent information of the current driving status of autonomously operated main motor vehicles; The uncertainties affecting the trajectory of the main vehicle are divided into expected uncertainties and unexpected uncertainties; The anticipated uncertainty in the first operational branch is calculated by forming an attention zone based on the lane identification portion, which may potentially conflict with the planned route of the main motor vehicle. The unexpected uncertainty in the second operational branch is determined by calculating the anomalous score of any other vehicle in the area surrounding the main motor vehicle in a lane that may potentially conflict with the planned route of the main motor vehicle. The unexpected uncertainty is defined along the axis of two opposing behaviors: including a preference for speed and a reward-seeking behavior in response to nearby agents, and a risk aversion including a willingness to wait for the uncertainty to decrease from a first uncertainty value to a second value below the first uncertainty value before continuing the expected movement of the main motor vehicle. If the unexpected uncertainty reaches or exceeds a predetermined threshold, the trajectory operation signal determined for the expected uncertainty is adjusted. A predetermined threshold is set so that the attention area is only used to calculate savings when the level of unexpected uncertainty is below the predetermined threshold; Attention zones are applied to reduce the computational cost required to make trajectory decisions, wherein trajectory data of vehicles within each attention zone that is defined as a high attention zone are used to determine when to perform a maneuver, and trajectory data of vehicles outside the attention zone are excluded. and Based on the adjusted trajectory operation signals, the main motor vehicle can move in a manner similar to that of a human. Smooth driving in traffic situations involving both humans and autonomous vehicles, wherein the human-like behavior of the main vehicle's movement is simulated after neural modulation processes of human cognition, including reward-seeking behavior and risk aversion, and wherein the human driver in the vicinity of the main vehicle is able to predict and anticipate the human-like movement of the main vehicle.

2. The method according to claim 1, further comprising: Lane information is represented as coordinates, heading, and speed limit.

3. The method according to claim 2, further comprising: When retrieving map information and agent information for the current driving state, the coordinates, heading, and speed of agents, including other vehicles in the surrounding area of ​​the main motor vehicle, are displayed.

4. The method according to claim 3, further comprising: The attention area is applied to identify agents to be filtered out, including any vehicle in the surrounding area of ​​the main vehicle outside the attention area, such that the agents to be filtered out are not used for further processing.

5. The method according to claim 1, further comprising: Generate potential trajectory branches for the trajectory of the main motor vehicle.

6. The method according to claim 5, further comprising: Select the optimal or "best" trajectory branch from the potential trajectory branches to be executed by the main motor vehicle.

7. The method according to claim 1, further comprising: The speed and heading of other vehicles in the area surrounding the main motor vehicle are compared with the individual expected speed and heading of other vehicles in their current lane positions.

8. The method according to claim 7, further comprising: The aggregate score for the unexpected uncertainty is obtained by averaging the anomaly scores of any other vehicles.

9. The method according to claim 1, further comprising: A first adjustment signal is generated, which is applied to temporarily disable the filtering of the attention zone of any other vehicle during the adjustment trajectory operation signal.

10. The method according to claim 1, further comprising: The second operation branch is executed in parallel with the first operation branch.

Images